Target repositional accuracy and PTV margin verification using three-dimensional cone-beam computed tomography (CBCT) in stereotactic body radiotherapy (SBRT) of lung cancers

Abstract

The purpose of this study was to assess target repositional accuracy with respect to the bony structures using daily CBCT, and to validate the planning target volume (PTV) margin used in the lung SBRT. All patients underwent 4D CT scanning in preparation for lung SBRT. The internal target volume (ITV) was outlined from the reconstructed 4D data using the maximum-intensity projection (MIP) algorithm. A 6 mm margin was added to the ITV to create the PTV. Conformal treatment planning was performed on the helical images, to which the MIP images were fused. Prior to each treatment, CBCT was taken after a patient was set up in the treatment position. The CBCT images were fused with the simulation CT based on the bony anatomy, in order to derive setup errors and separate them from the tumor repositional errors. The treating physician then checked and modified the alignment based on target relocalization within the PTV. The shifts determined in such a method were recorded and the subtractions of these shifts with respect to the corresponding setup errors were defined as the target relocalization accuracy. Our study of 36 consecutive patients, treating 38 targets for a total of 153 fractions shows that, after setup error correction, the target repositional accuracy followed a normal distribution with the mean values close to 0 in all directions, and standard deviations of 0.25 cm in A-P, 0.24 cm in Lat, and 0.28 cm in S-I directions, respectively. The probability of having the shifts ? 0.6 cm is less than 0.8% in A-P, 0.6% in Lat, and 1.7 % in S-I directions. For the patient population studied, the target centroid position relative to the bony structures changed minimally from day to day. This demonstrated that the PTV margin that is designed on the MIP image-based ITV was adequate for lung SBRT.

Figure 1. An example of image fusion between daily cone‐beam images and the simulation images: (a) after bony alignment but before soft tissue alignment; (b) after soft tissue alignment.

Figure 2. Setup errors in A–P (a), Lat ((b), and S–I (c) directions for 153 fractions. Data were derived based on bony landmark alignments.

Figure 3. The distributions of the setup errors of all three major directions.

Figure 4. Average setup errors per patient/per target sorted based on the tumor locations (LLL, LUL, RLL, and RUL): (a), (b), and (c) correspond to the setup errors in the Lat, A–P, and S–I directions, respectively.

Figure 5. Target relocalization uncertainties along the three major directions with respect to the patient bony structure, which are defined as the additional shifts that the physician has made in order to align the target within the PTV.

Figure 6. Distributions of the target relocalization uncertainties along the three major directions.

Figure 7. Average target relocalization uncertainties per patient/per target sorted based on the tumor locations (LLL, LUL, RLL, and RUL): (a), (b), and (c) correspond to the target uncertainties in the Lat, A–P, and S–I directions, respectively.